Module 03 / 08
Retrieval & RAG
Learn semantic retrieval, chunking, and grounded generation as one system.
The point
Retrieval and generation fail together in real products. Treating them as one layer makes it easier to debug relevance, citations, and answer quality.
Start with
Read next
Understanding check
You should be able to…
- explain semantic search and cosine similarity
- choose a chunking strategy for a document type
- inspect retrieved context before blaming the model
- separate retrieval quality from answer quality
- create a small eval set for grounded answers
Practice with an agent
Add cited answers to your product
beginner · starter repository, model API, prepared documents
Make
a small retrieval slice that answers from a prepared document set and exposes the retrieved chunks
You know it works when
answer five questions with citations and identify whether each failure came from retrieval or generation
Go deeper by building
Build a cited knowledge assistant
intermediate · next.js or python, embeddings API, sqlite or a vector store, model API
Make
a local app that searches a document folder and answers with citations
You know it works when
include retrieval logs, citations, and ten grounded-answer evals with pass/fail notes